Coordination in cooperative distributed systems

A cooperative distributed systems approach is an appropriate design paradigm for many applications, such as manufacturing control and network management. A fundamental problem of this paradigm is how to manage interdependencies among the cooperating entities of the system to achieve their goals. The solution to this interdependency problem is commonly known as coordination. This thesis presents a novel way of understanding and analyzing coordination in cooperative distributed systems.;Coordination is looked at in terms of structure and mechanism. This thesis dwells mostly on the mechanism, with brief attention to the structure. A mechanism is a composition of a set of decision points and devices directed to deal with different aspects of the interdependency problem. To deal with the dynamic and uncertain characteristics of the environment, the concepts of intelligence and rationality are introduced. Furthermore, to cope with time-bounded situations, rationality is employed at meta-reasoning level and action level.;This thesis provides a bridge between theoretical analysis and application environments for coordination by developing an agent-based model for cooperative distributed systems. The architecture of the model is based on the agent's mental state regarding its reasoning about achieving a goal, and consists of four components: problem-solver, pre-interaction, interaction, and execution.;The feasibility of the proposed model has been demonstrated by applying it to a distributed computing system handling the routing and the control of information flow in a distributed fashion.